R/mean_and_ci.R

In lorenc5/eiCompare: Compares EI, Goodman, RxC Estimates

mean_and_ci <- function(cbind_dat, ci=c(.025, .975)){
  # Add votes together
  total <- apply(cbind_dat, 1, sum)
  # Create vote filler
  v_fill <- matrix(NA, nrow=nrow(cbind_dat), ncol=ncol(cbind_dat))
  # Create mean, CI filler
  qtile <- matrix(NA, nrow=ncol(cbind_dat), ncol=3)
  # Create percents for every vote column for 1 racial group
  for (j in 1:ncol(v_fill)) {
    # Divide each vote by the total
    v_fill[,j] <- cbind_dat[,j] / total
    # Mean, 95% confidence interval
    qtile[j,]<- c(mean(v_fill[,j]), quantile(v_fill[,j], ci) ) 
  }
  # Label Output #
  row.names(qtile) <- colnames(cbind_dat)
  colnames(qtile) <- c("Mean", "2.5", "97.5")
  # Return Mean, CI table
  return(qtile)
}